[chromium] Move to USE(LAZY_NATIVE_CURSOR)

• Component: Platform
• Status: RESOLVED FIXED

[Rick Byers]

The 'Cursor' type is an opaque object except on ports that have USE(LAZY_NATIVE_CURSOR).  It looks like most/all ports other than chromium use this.  On chromium it would make the Cursor object redundant with PlatformCursor (since they'd both just be a container for the values due to chromium's multi-process architecture - we don't manipulate real OS cursors from the renderer process), so doesn't necessarily add value.  However it looks like it would have testability benefits to unify testing across ports.

I'm going to try enabling this on chromium, and then we can decide if it's worthwhile.

[Rick Byers]

Created attachment 173047 [details]
Patch

[Rick Byers]

This actually turned out very nicely.  We can simplify things by removing PlatformCursor on chromium entirely - USE(LAZY_NATIVE_CURSOR) gives us effectively the same thing but in a simple fashion that is unified with most other ports: a copy of the cursor information that will eventually be used to construct a real OS cursor (but in Chromium's case this isn't done until the data is passed to the browser process). 

However I believe this has the side effective of changing at least some cursor behavior, in particular fixing bug 100059.  So I don't think we should make this change until there are tests in place.  I'll work on landing bug 100550 for the other ports that already have USE(LAZY_NATIVE_CURSOR) first.

[Adam Barth]

Comment on attachment 173047 [details]
Patch

Looks great!

[Rick Byers]

Created attachment 173176 [details]
Combined with fix for 100550 for EWS

[Early Warning System Bot]

Comment on attachment 173176 [details]
Combined with fix for 100550 for EWS

Attachment 173176 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/14771465

[Early Warning System Bot]

Comment on attachment 173176 [details]
Combined with fix for 100550 for EWS

Attachment 173176 [details] did not pass qt-wk2-ews (qt):
Output: http://queues.webkit.org/results/14775347

[Build Bot]

Comment on attachment 173176 [details]
Combined with fix for 100550 for EWS

Attachment 173176 [details] did not pass win-ews (win):
Output: http://queues.webkit.org/results/14779175

[Build Bot]

Comment on attachment 173176 [details]
Combined with fix for 100550 for EWS

Attachment 173176 [details] did not pass mac-ews (mac):
Output: http://queues.webkit.org/results/14763871

[kov's GTK+ EWS bot]

Comment on attachment 173176 [details]
Combined with fix for 100550 for EWS

Attachment 173176 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/14780102

[Rick Byers]

Created attachment 173375 [details]
Merge with ToT

[Rick Byers]

Created attachment 173376 [details]
Combined with fix for 100550 for EWS

[Adam Barth]

Comment on attachment 173375 [details]
Merge with ToT

This looks great.

[WebKit Review Bot]

Comment on attachment 173376 [details]
Combined with fix for 100550 for EWS

Attachment 173376 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/14790306

[Rick Byers]

Created attachment 173425 [details]
Patch for landing

[Rick Byers]

Created attachment 173429 [details]
Fix bad merge - actually re-enable the test in chromium

[Rick Byers]

Created attachment 173430 [details]
Patch for landing

[WebKit Review Bot]

Comment on attachment 173430 [details]
Patch for landing

Clearing flags on attachment: 173430

Committed r134149: <http://trac.webkit.org/changeset/134149>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.